How would you design parallel grep for huge JSONL files?

I need to scan very large JSONL files efficiently and am considering a parallel grep-style approach over line-delimited text. Would love to hear how you would design it.